slade benner silva ieee proceedings
TRANSCRIPT
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
INV ITEDP A P E R
Goldstone Solar System RadarObservatory: Earth-BasedPlanetary Mission Support andUnique Science Results
By Martin A. Slade, Member IEEE, Lance A. M. Benner, and Arnold Silva
ABSTRACT | The Goldstone Solar System Radar (GSSR) facility is
the only fully steerable radar in the world for high-resolution
ranging and imaging of planetary and small-body targets. These
observations provide information on surface characteristics,
orbits, rotations, and polar ices for a wide variety of solar system
objects. The resulting data are used not just for scientific studies
of these objects, but also for frequent support of the National
Aeronautics and Space Administration (NASA) flight projects,
including many solar system exploration missions over the last
three decades. For example, the GSSR has contributed to the
Mars Exploration Rovers (MERs), Cassini, Hayabusa (MUSES-C),
MESSENGER, NEAR, SOHO recovery, Mars Pathfinder, Lunar
Prospector, Clementine, Magellan, and Viking. Other recent
examples include measurement of lunar topography at high
resolution near the lunar south pole, which is of particular
interest concerning the impact site of the Lunar Crater Observa-
tion and Sensing Satellite (LCROSS) mission, and the character-
ization and orbit refinement of near-Earth asteroids, both for
asteroid impact hazard mitigation and for identification of
potential targets for future spacecraft missions. We also present
important radar scientific results including near-Earth object
(NEO) radar imaging of especially interesting objects, and the
results from high accuracy determination of Mercury rotation via
radar speckle displacement (RSD).
KEYWORDS | Chirp modulation; CW radar; delay estimation; HF
radar; planets
I . INTRODUCTION
There are two powerful ground-based radars in the world
capable of investigating solar system objects: the National
Aeronautics and Space Administration (NASA) Goldstone
Solar System Radar (GSSR) and the National Science
Foundation (NSF) Arecibo Observatory. NASA derives
both scientific and programmatic benefits from their use,
including investigations of the Moon, planets, and othersolar system bodies, including near-Earth objects (NEOs).
These studies provide images and topography, and
information on surface characteristics, shapes, composi-
tion, orbits, rotations, and distribution of liquids or ices on
or below their surfaces. These radars also provide frequent
and singular support of NASA missions. The capabilities of
these two instruments are different and often provide
complementary data, so the combination of results fromboth radars has often proven to be much more valuable to
NASA than each individually, sometimes yielding valuable
information that neither could obtain alone. The Arecibo
radar is located in Puerto Rico. For a detailed review of the
elements of the Arecibo radar, see [1]. This paper will
focus on the GSSR.
Manuscript received March 1, 2010; revised August 27, 2010; accepted September 9,
2010. This work was carried out at the Jet Propulsion Laboratory, a division of
the California Institute of Technology, under contract with the National Aeronautics
and Space Administration (NASA).
M. A. Slade and L. A. M. Benner are with the Jet Propulsion Laboratory, California
Institute of Technology, Pasadena, CA 91109 USA (e-mail: [email protected]).
A. Silva was with the Jet Propulsion Laboratory, California Institute of Technology,
Pasadena, CA 91109 USA. He is now with the Aerospace Corporation, El Segundo,
CA 90245 USA.
Digital Object Identifier: 10.1109/JPROC.2010.2081650
| Proceedings of the IEEE 10018-9219/$26.00 �2010 IEEE
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
II . THE GOLDSTONE RADARFACILITY DESCRIPTION
A. OverviewThe NASA/JPL Deep Space Network (DSN) supports
the GSSR. The GSSR consists of a �500-kW transmitter
with center frequency at X-band (8560 MHz or 3.5-cm
wavelength) on the 70-m antenna at the Goldstone DSNsite (near Barstow, CA). See Fig. 1. GSSR has illuminated
targets for reception in multistation mode for receivers at
Arecibo Observatory, the very large array (VLA), the very
long baseline array (VLBA), Evpatoria (Ukraine), Medicina
(Italy), Usuda and Kashima (Japan), as well as five 34-m
DSN telescopes also at the Goldstone site. It has performed
two-station observations with the 100-m Green BankTelescope (GBT, Green Bank, WV) for planetary targets
Mercury and Venus. (See Appendix I for the locations of
the United States telescopes discussed above.)
In the text below, we will introduce the concepts of
continuous-wave (CW) radar, radar delay-Doppler map-
ping, the use of radar interferometry to measure topogra-
phy, the imaging of NEOs by radar, and various other
scientific applications of radar techniques by the GSSR.
B. GSSR ComponentsThe GSSR capability is built around an X-band (8.6 GHz)
high-power transmission capability, and can be broken
down into the transmitter, receiver, and data-acquisition
subsystems. A conceptual diagram of the interfaces and flow
among the subsystems is shown in Fig. 2. (A detailed set of
actual block diagrams is shown in Appendix II.) Thetransmitter subsystem consists of two CW Communications
Power Industries (CPI) klystrons amplifiers (Fig. 3), along
with their associated power system and cooling system. The
two Model VKX-7864 250 kW klystrons currently in use are
a matched pair that were put in the BX-band-K-band-
Research[ cone or XKR cone in September 2010. The radar
encoding on the transmitted signal is achieved by mixing the
signal from an exciter in the transmitter assembly withsignals from various waveform generators. The klystrons
reside up inside the BXKR cone[ on the steerable 70-m
antenna, below the subreflector at the prime focus. The
high-power radar signal is transmitted out its horn, and a
Bquasi-optical[ transmit–receive (T/R) switch, which is a
GSSR component especially developed for NEO applica-
tions, allows for receiving with low latency. Because the
received signal is acquired through a separate hornconnected to a high electron mobility transistor (HEMT)
Fig. 1. (Left) 70-m DSS-14 antenna at the Goldstone DSN complex. Note
the three cones below the subreflector on the antenna. (Right) The
XKR-cone has the GSSR transmit and receive horns and the
transmit/receive (T/R) switch (black arrow). The inset image
shows the T/R switch in the receive position.
Fig. 2. Diagram of GSSR subsystem (S/S) interfaces. The dual-channel receiving LNA and data processing chain is capable of recording
right-circularly-polarized and left-circularly-polarized (RCP and LCP) signals simultaneously. The T/R switches between the
two feed horns, one horn from the transmitter, the other to the LNA. (See Fig. 1.)
Slade et al. : Goldstone Solar System Radar Observatory
2 Proceedings of the IEEE |
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
low noise amplifier (LNA) also located inside the XKR cone,
observing nearby NEOs requires switching the transmitter
on and off, but also switching between the transmit and
receive feeds up on the XKR cone (Fig. 1).The data acquisition subsystem (DAS) is critical for
conversion of the analog signals from the receiver to digital
data (A/D converters) that is recorded to disks for offline
data analysis. (The receiver performs appropriate signal
processing of the LNA’s output, depending on the goals of
the particular observation.) We include the software that is
used to command and control the radar in the DAS. Parts
of that software use ephemeris information to drive theprogrammable oscillators (POs) that are an important part
of the GSSR system. The POs take input from the DSN’s
stable master frequency reference. (See Appendix II for a
more complete description of the POs.)
III . RADAR DELAY-DOPPLER MAPPING
The central feature of delay-Doppler mapping is the
separation of the backscattered signal from various resolu-
tion cells based on their distance (range) and line-of-sight
velocity with respect to the observer. The radar observablesare the round-trip time delay and Doppler frequency shift,
with the Doppler shift arising from orbital and rotational
motions of the Earth and the target body. As illustrated in
Fig. 4, regions at a given time delay (range) from the radar
are planes perpendicular to the line-of-sight direction i. A
region with a constant Doppler frequency are planes parallel
to a plane containing the line-of-sight direction i and to the
apparent rotation axis k. Analyzing the radar echoes in time
delay and Doppler frequency thus measures the radar
backscatter from different parts of a target.As shown in Fig. 4, resolution cells on the surface can
have the same delay and Doppler values, such that their
echo powers combine in a delay-Doppler map (e.g., A and
B in Fig. 4) and the two physical hemispheres collapse into
one. This condition is called the BNorth–South ambiguity.[An example of a delay-Doppler map is shown in Fig. 5 for
Fig. 3. GSSR XKR cone and a CPI 250-kW klystron amplifier. The
horn leaving the top of the cone in the cone axis is the transmitter
horn, and the one to its left is attached to the LNA. (Varian, Inc.
merged with several other companies in 2004 to form CPI. The LNA
in this figure is labeled ‘‘XKR Maser,’’ which has been replaced in 2008
by an X-band high electron mobility transistor or X-HEMT LNA.)
Fig. 4. Delay-Doppler mapping geometry for a spherical body. Planes
of constant time delay from the radar are planes perpendicular to the
line-of-sight direction i. Planes of constant Doppler frequency are
planes parallel to a plane containing the line-of-sight direction i and to
the apparent rotation axis k. (Figure adapted from [2].)
Fig. 5. Delay-Doppler map of Mercury obtained on June 16, 2000,
at Arecibo. The leading edge of the radar echo is the parabolic locus
that is brightest at the top. The North Pole craters that are filled
with radar-bright material (putatively water ice) are at the bottom
center. The delay resolution is 3 km. (From [3].)
Slade et al.: Goldstone Solar System Radar Observatory
| Proceedings of the IEEE 3
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
the planet Mercury as seen from Arecibo [3]. The North–
South ambiguity is not important here since the terrain
to the south of the area imaged is very radar dark. Fig. 6
focuses on the North Polar craters, which are seen at the
bottom center of Fig. 5, but imaged with Goldstone data
in planetary coordinates from data collected in June andJuly 2001 [4].
The delay-Doppler technique makes use of the relative
motions between imaged area points to create the Doppler
resolution. At high spatial resolution, regions at the edges
of the imaged area will Bsmear[ due to migration in delay
and Doppler cells not accounted for in simple delay-
Doppler imagery [5]. Either focused range-Doppler
processing or polar-format spotlight-mode processing canbe used for better image quality. See discussions of the
former method for high-resolution imaging in [5] and the
latter method in [6].
IV. RADAR INTERFEROMETRY
Two-station Earth-based radar interferometry can obtain
detailed topographic maps of the surfaces of planetary
bodies. The technique is similar to terrestrial topographicmapping from interferometric synthetic aperture radar
(SAR) observations and was first described in [7] and [8].
Since the original formulation of the technique for lunar
mapping, radar interferometry has been applied to
terrestrial remote sensing [9]. The technique has been
used with considerable success in a variety of geophysical
applications [10]. The description here of delay-Doppler
mapping and of the basic concept underlying radarinterferometry closely follows the approach presented in
[7] and [11]. Consider an Earth-based radar system with
two receiving stations forming an interferometer. The
distance between the two interferometric receiving
antennas, the baseline B, is assumed to be much smaller
than the range R to the target body under consideration. In
the usual radar experiment, the experimenter records
radar echoes in both amplitude and phase in order toextract information about the material properties and
location of the scattering surface. When a radar interfer-
ometer is used, additional information about the location
of the reflecting surfaces can be obtained because
incoming wavefronts have different arrival times at the
two receiving antennas. The time delay depends on the
orientation of the incident radiation and can be related to
the interferometric phase (that is, the phase differencebetween the two radar echoes). As an example, a 2� phase
change is recorded by the interferometer when the
orientation of incoming radio waves changes by �=Bp,
where � is the wavelength and Bp is the baseline projected
on the plane perpendicular to the line of sight. At a
distance R, this angular difference corresponds to a fringespacing s, which is given by
s ¼ R�
BP: (1)
Topographic changes are then related to the measured
interferometric phase because the location of reflection
elements in the fringe pattern of the interferometer is
elevation dependent. Height deviations �z parallel to the
projected baseline produce phase changes �� for the
interferometer in proportion to
�z
s¼ ��
2�: (2)
Combining (1) and (2), we obtain
�z ¼ R���
ð2�BPÞ: (3)
The above basic concepts are used in the topographic
mapping for the Moon presented in [12]. See also
Section VI.
V. NEAR-EARTH OBJECTS
Ground-based radar is the most powerful astronomical
technique for characterizing NEOs and refining their
Fig. 6. Goldstone image of the North Pole of Mercury obtained in 2001
[4]. The delay-Doppler data were mapped to planetary coordinates
in a polar orthographic projection. The resolution is 6 km by 6 km.
Slade et al. : Goldstone Solar System Radar Observatory
4 Proceedings of the IEEE |
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
orbits. While NEOs appear as unresolved points through
ground-based optical telescopes, the Arecibo and Gold-
stone radars can image NEOs with resolution as fine as
several meters. Radar has produced the best physical
characterizations of potentially hazardous NEOs as large as
a kilometer and the best images yet of binary small body
systems of NEOs.
Radar echoes from NEOs have revealed both stony andmetallic objects, featureless spheroids, and shapes that are
elongated and irregular, objects that must be monolithic
chunks of solid rock, and objects that must be unconsol-
idated Brubble piles.[ See Fig. 7 for examples of delay-
Doppler radar images of NEOs. (More elaborate modeling
of the shapes of NEOs can be performed using the
techniques pioneered by Hudson [13] and more recently
updated and documented by Magri [14], [15].)
A. Nonprincipal Axis RotatorsRadar images of 1999 JM8 obtained in 1999 [16] reveal
an asymmetric, irregularly shaped, �7-km-wide object
with pronounced topography, prominent kilometer-sized
facets, and numerous crater-like features. 1999 JM8 is a
nonprincipal axis rotator with a periodicity of order oneweek.
Delay-Doppler images of the Earth-crossing asteroid
4179 Toutatis achieved resolutions as fine as 125 ns (19 m
in range) and 8.3 mHz (0.15 mm/s in radial velocity) and
placed hundreds to thousands of pixels on the asteroid,
which is 4.5 km long and 2.4 km wide, and heavily cratered
[20]. Radar observations from 1992, 1996, and 2004 reveal
Toutatis to be in an extremely slow, nonprincipal-axisrotation state with two main periods of 5.4 and 7.3 days.
The best shape model has an areal resolution of
approximately (34 m)2 [21].
B. Orbit RefinementRadar is invaluable for refining the orbits of potentially
hazardous NEOs. Range-Doppler shift measurementsprovide line-of-sight positional astrometry with precision
as fine as 4 m in range and 1 mm/s in velocity, with a
fractional precision typically 100–1000 times finer than
with typical optical measurements [22]. Radar reconnais-
sance can add decades to centuries for the intervals over
which we can predict close Earth approaches and
dramatically refines collision probability estimates based
on optical astrometry alone, especially for objects passing
the Earth for the first time.
The NEO population also provides the most accessible
bodies for robotic and human spacecraft missions.
Spacecraft operations to these primordial bodies can be
greatly aided by radar-derived physical models and
precision trajectories of the mission targets, reducing the
mission cost, decreasing the complexity in the design ofthe spacecraft, and improving the odds of a successful
mission, be it sample return or hazard mitigation.
C. Binary Asteroid SystemsWith binary NEO 1999 KW4, Ostro et al. [23]
discovered a suite of exotic phenomena, such as a librating
secondary, an oblate shape, and a pronounced equatorialridge rotating at nearly escape velocity (see Fig. 8).
Other multiple-body NEOs observed at Goldstone
include contact binary 2007 VD12 in November 2007,
and a triple body object 1994 CC first identified as such in
June 2009.
D. Higher Range Resolution: Observations in 2010NEO observations by the GSSR up to 2009 have used
only binary phase coded (BPC) waveforms to deliver the
range resolution (also known as PN coded) (see, for
example, [24]). The best range resolution at Goldstone
using BPC waveforms has been limited to 18.75-m
resolution (1/8 �s), since any finer resolution violates
the frequency license. (GSSR’s license from the National
Telecommunications and Information Administration(NTIA) is between 8500 and 8620 MHz.)
However, a change in waveform can achieve higher
resolution without violating the GSSR license: the linear
frequency modulation (LFM or chirp) waveform, which is
widely used in synthetic aperture radar (SAR) applications
(see, e.g., [25]). (Using a chirp waveform for GSSR ranging
is based on an informal report [26].) An internal JPL study
[27] found that the practical bandwidth limit for thecurrent Goldstone klystrons is 40 MHz when using chirp
waveforms, which leads to an imaging system with 3.75-m
resolutionVan improvement of a factor of five. (Note that
this is twice as fine as the best range resolution of Arecibo’s
S-band planetary radar.) A sequence of images obtained in
January 2010 with 3.75-m range resolution of NEO 2010
AL30 is shown in Fig. 9.
Fig. 7. Goldstone NEO images obtained via delay-Doppler mapping ([16]–[19], respectively).
Slade et al.: Goldstone Solar System Radar Observatory
| Proceedings of the IEEE 5
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
By using both a broadband transmitter and a chirp
waveform with 150-MHz bandwidth, the range resolution
could be as fine as 1 m. With such fine scale imaging, this
ground-based radar observatory could produce Bspacecraft-
encounter-quality[ images of NEOs several times per year,
providing an extremely low-cost means for NASA to
explore a sizable fraction of the NEO population.
VI. THE MOON
NASA is currently planning and performing an ambitious
set of lunar robotic missions to determine the topographyand gravity field of the Moon to very high precision.
The south pole of the Moon has attracted much recent
attention due to the possibility of significant amounts of
frozen water being trapped in permanently shadowed
regions in craters in that region. Key to understanding of
the possibility of the frozen volatiles at the poles of the
Moon is detailed knowledge of the lunar polar topography.
Radar inteferometry from the Goldstone complex has an
ongoing history of providing the most accurate such lunar
topographic maps available, with ever-increasing detail
revealed as the radar mapping techniques evolve. From
topography with 150-m pixels in 1997 [11] for aiding theLunar Prospector and Clementine Missions to 20-m pixels
in 2007, the Goldstone radar was essential. Fig. 10 shows
the GSSR imaging of the Lunar North Pole (center) in
2009 compared with two spacecraft imagesVa composite
of the SAR imaging by the Indian lunar orbiter Chandrayan
and the optical imaging by the 1994 United States lunar
orbiter ClementineVof the same region.
In 2009, the 5-m pixel imaging of the Moon madeessential contributions to planning and postimpact analysis
for the Lunar Crater Observation and Sensing Satellite
Fig. 9. Goldstone delay-Doppler imagery of NEO 2010 AL30 on January 13, 2010, was using the highest resolution chirp waveform. Each
frame represents 30 s of integration. Range increases downward; 1.875 m separate each row of pixels. Doppler increases to the right; each
pixel column is separated by 0.5 Hz in each frame of the montage. The vertical drift of the echo in the images is due to uncertainties in the
ephemeris. The panel on the right shows the trajectory of 2010 AL30 in the vicinity of the Earth and the Moon (orbital image from [28]).
Fig. 8. Joint Goldstone/Arecibo radar imaging of asteroid 1999 KW4. Arecibo observations allowed precision modeling of extreme oblateness
of primary object (left) while Goldstone observations revealed details of the orbit of secondary object (right) [23].
Slade et al. : Goldstone Solar System Radar Observatory
6 Proceedings of the IEEE |
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
(LCROSS) mission [12]. The Goldstone radar has given
good return on investment by timely delivery of high-
resolution Lunar maps, with height resolution improving
from 50 m in 1997 [11] to 3 m [root mean square (rms)] in2009–2010.
Even with topographic mapping by a variety of orbiting
platforms [e.g., Lunar Reconnaissance Orbiter (LRO)], the
topographic maps generated by the Goldstone radar will
remain the highest resolution contiguous maps over a large
area for the foreseeable future. Although missions such as
LRO will generate topographic measurements at points
over much of the lunar surface with a vertical accuracy onthe order of 1 m, the horizontal separation of these vertical
measurements will be less than 20 m only at the poles. At
the lunar equator, the horizontal separation will be �1 km
[29]. Optical stereogrammetry-derived digital elevation
models (DEMs) will provide high-resolution topographic
maps of small selected areas, but lack the large area
coverage (105 km2) of a typical Goldstone radar topo-
graphic map of portions of the lunar surface.
VII. PLANETARY OBSERVATIONS ANDNASA MISSION SUPPORT
Both the Goldstone and Arecibo radars have a long and
continuing history of exploring the deeper solar system,
both for basic solar system science and for NASA missionsupport. Present and past uses of these radars are briefly
described below, starting with the inner solar system and
working outward.
Mercury has long been a target for ground-based radar
observations. The Shapiro Bfourth[ test of general
relativity was an early (1964) triumph. Routine radar
ranging to Mercury has contributed to knowledge of its
orbit for spacecraft to Mercury mission navigation,
including Mariner 10 and the current MESSENGER
mission. The Goldstone radar is contributing to the
MESSENGER mission by precisely measuring the forcedlongitude librations of Mercury. The technique of obser-
vation has been dubbed Bradar speckle displacement[effect (RSD effect). In brief, a CW radar reflection from
the solid surface of Mercury exhibits spatial irregularities
in its wavefront caused by the constructive and destructive
interference of waves scattered by the irregular surface of
Mercury. The corrugations in the wavefront, also called
speckles, are tied to the rotation of Mercury, and willsometimes sweep over the Earth’s surface as Mercury
rotates. When the trajectory of the speckles is parallel to
the baseline between two antennas, the radio signals at the
two receiving telescopes show a high degree of correlation
after allowing for the time lag from the finite speed of the
speckle motion. (Fortunately, the relative locations of the
Earth and Mercury and the orientation of their rotation
poles are known well enough that it is possible to predictthe dates and times of speckle trajectories traversing the
baseline of the antennas.) The amount of lag delay that
maximizes the correlation of the echoes at the two
receiving telescopes can be shown [30] to be directly
related to the magnitude and orientation of the spin vector
of the planet. The large measured libration amplitudes (see
Fig. 11) from Goldstone and the GBT observations (larger
than if the entire planet responded to the solar torque)demonstrated that the Mercury mantle must be decoupled
from the deeper interior by a molten layer. The pole
orientation of Mercury is also being determined with
unprecedented accuracy from the same set of radar
observations (the discussion above is based on [31] and
the references therein).
Fig. 10. Example of 2009 capability of Goldstone radar at the lunar North Pole compared with two images of the same region
from lunar orbiters. The pixel resolution is 5 m by 5 m in the bottom image [12].
Slade et al.: Goldstone Solar System Radar Observatory
| Proceedings of the IEEE 7
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
The BRSD[ technique is also being applied by
Margot et al. to Venus using Goldstone and the GBT to
determine the interaction of the Venus surface with thevery dense atmosphere, as well as refinement of the spin
rate and pole direction.
Both the Goldstone and Arecibo radars mapped
portions of Venus prior to the Magellan mission to
determine the Venusian spin rate and pole direction,
which allowed the Magellan to enter an optimal mapping
orbit around Venus. Both radar observatories determined
these quantities from different sets of features on theVenus surface. (See [32] and [33].) The separate
determinations agreed very well, and a combination of
all the time histories of the Venus radar features led to achoice for the Magellan orbit that worked perfectly.
Mars also has been a frequent target of ground-based
radars. In 1971 and 1973, an intense ranging campaign of
the Mars equatorial regions led to topographic maps,
which were the standard reference for many years [34].
Since then, every Mars landing mission, including Viking,
Pathfinder, the Mars Exploration Rovers (MERs), and
Phoenix, has used the ground-based radars to help winnowdown the set of scientifically interesting landing sites by
measurement of their surface roughness (e.g., [35]–[38]).
Since Pathfinder and the MERs were dependent on a solid-
surface-sensing radar for final descent, a key contribution
of the Goldstone radar was in validating the radar
reflectivity of the various sites in addition to the roughness
determinations.
The Cassini mission to Saturn carried the European-built Huygens probe to land on Titan, the largest moon of
Saturn. The Huygens probe descended into the thick
atmosphere of Titan, which some scientists hypothesized
was a by-product of an ocean of hydrocarbons that was
kilometers deep. Joint observations by the Goldstone radar
and the VLA determined that these fears were exaggerated,
and that continent-sized regions of ice protruded through
the ocean [39]. The Arecibo radar also observed Titan at13 cm in 2001 and 2002, and obtained specular echoes
consistent with reflection from areas of liquid hydro-
carbons [40]. The radar on-board the Cassini spacecraft
performed flyby radar imaging of Titan. After the Huygens
Probe descent to Titan’s surface, the Cassini radar
determined that hydrocarbon (ethane mixed with meth-
ane) lakes could be found, but no very large seas, much less
a ubiquitous kilometers-deep ocean.In 1999, the Goldstone radar played an unusual role in
support of the ESA/NASA SOHO mission, actually resulting
in helping to save the spacecraft [41]. The SOHO solar
observatory, in an L1-halo orbit near the Earth, accidentally
used a large part of its attitude control thruster fuel, which
left the spacecraft spinning at a rate that made telemetry
contact useless. Ground-based radar observations in which
Arecibo-transmitted and Goldstone-received CW radar
Fig. 11. Mercury spin rate deviations from the resonant rate of 3/2
times the mean orbital frequency. Observed data points and their error
bars are shown in black. The red line shows a numerical integration of
the Sun’s torque on Mercury, which is fitted to the data. This fit
estimates three parameters, which in effect include the 88-day forced
libration and a 12-year free libration. (Figure from [31].)
Fig. 12. The chirp ranging system increases the GSSR ranging resolution by a factor of five, from 18.75 to 3.75 m. This figure shows the significance
of a factor of five improved resolution for NEO 1998 CS1 from 75 m per row at the top to 15 m per row at the bottom.
Slade et al. : Goldstone Solar System Radar Observatory
8 Proceedings of the IEEE |
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
returns at S-band from SOHO determined the rate at which
the SOHO spacecraft was spinning. Moreover, the radar
echoes showed that the spin rate was low enough that a
fraction of the remaining thruster propellant could bring
SOHO back to rotating at a Sun-synchronous rate. As of 2010,
SOHO (which cost about a billion Euros) was still operating.
One additional use of the Goldstone radar is itsparticipation in a NASA program of monitoring orbital
debris around the Earth. The Goldstone radar provides
unique information on clouds of extremely small particles,
ranging from about 1–10 mm, orbiting between altitudes of
about 400–2500 km. [42]. Even particles of this small size,
moving at extremely high velocities, can pose safety
hazards to both human and robotic spacecraft. These radar
observations continue to be important to astronautsmoving on the exterior of the International Space Station.
VIII . FUTURE DIRECTIONS FORTHE GSSR
As mentioned above, the resolution of NEO imaging at the
Goldstone radar could be improved to significantly finer
than the current 3.75-m range resolution. Even though this
is a factor of five higher than the previous best rangeresolution at Goldstone (see Fig. 12 for a pictorial version
of what a factor of five can gain), this could be improved to
about 1-m range resolution with no difficulty in principle.
The fundamental limitation to the range resolution is the
bandwidth of the klystron power amplifiers through which
the Bchirp[ waveform passes. The spectrum allocation of
the Goldstone radar is from 8500 to 8620 MHz (120-MHz
Fig. 13. The locations of the VLBA telescopes, the GBT, Arecibo, Goldstone, and the VLA within the United States.
Fig. 14. The locations of radio telescopes at the Goldstone Deep Space
Communications Complex (GDSCC). North is towards the top. The scale
can be quickly understood from knowing that the baseline length
between Deep Space Station (DSS)-25 and DSS-13 is 13 km. Three
telescopes are nearly colocated at the ‘‘Apollo’’ complex: DSS-25,
DSS-24, and DSS-26. All of these radio telescopes have been used as
receivers by the GSSR.
Slade et al.: Goldstone Solar System Radar Observatory
| Proceedings of the IEEE 9
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
Fig. 15. (a) A block diagram of the GSSR hardware in the DSS-14 pedestal (immediately below the azimuth bearing), the so-called MOD III
area above the pedestal and below the elevation tracks. The colors are assigned as in Fig. 2. The area to the upper left shows telescopes
elsewhere in the Goldstone complex. Acronyms: FOT ¼ fiber optic transmitter; FOR ¼ fiber optic receiver; BVR ¼ block V Receiver
(DSN hardware); DRCV ¼ digital receiver (not shown is DRCV disk storage array); BPF ¼ band pass filter; PFS ¼ portable fast sampler;
TCT ¼ time code translator. (b) Continues from Fig. 15(a) as indicated there.
Slade et al. : Goldstone Solar System Radar Observatory
10 Proceedings of the IEEE |
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
bandwidth). The range resolution of the Goldstone radarcould be improved most simply by procuring power
amplifiers that would safely pass a broader bandwidth
chirp than the 40-MHz bandwidth amplifiers currently in
use. If power amplifiers could be obtained with 150-MHz
of usable bandwidth, then 1-m range resolution could be
achieved. However, to achieve 1-m resolution in cross
range (Doppler shift), longer coherent integration times
would be required. Only then could images with (1 m)2
pixel sizes be made.
Delay-Doppler radar NEO images with 3.75-m range
resolution will lead to dramatic improvement in our
knowledge of NEO sizes, shapes, and detailed surface
structure. Finer details for NEO surfaces will facilitate
much improved insight into NEO geology and collisional
evolution. The detection of asteroid satellites will be
significantly enhanced by the increase in resolution. Formany NEOs, the higher resolution will permit much finer
fractional precision Goldstone radar ranging astrometry
than was previously possible, which will improve long-term
orbital prediction for such objects.
IX. CONCLUSION
From the planet closest to the Sun to the icy moons ofJupiter and beyond to Saturn’s rings and Titan, the
Goldstone radar has performed extraordinarily valuable
science and contributed to nearly every NASA solar system
flight project, and is expected to continue doing so. The
NASA investment in this radar observatory has resulted in
significant enhancements (and frequently success-critical
observations) for NASA’s Flight Projects program, and has
greatly benefited NASA’s science investigations. h
APPENDIX IRADIO TELESCOPES IN THE UNITEDSTATES USED WITH THE GSSR
The telescopes to which the GSSR transmits (some
shown in Fig. 13) must have extremely accurate frequency
standards such as a cesium or a hydrogen maser Bclock.[
Commercial cesium clocks have been available since the1950s. Such Bclocks[ are essential for multistation radar
astronomy. Synchronization to universal time Bcoordinated[(UTC) has been greatly simplified by the global positioning
system (GPS), which can provide very accurate timing
signals with accuracy of about 50 ns for observatories around
the world. The Goldstone complex (Fig. 14) has frequency
and timing signals distributed to all telescopes by buried fiber
optical cable at sufficient depth to avoid daily heating effectschanging the fiber propagation speed. See [43] for the factors
that permit the Goldstone telescopes to operate as a very
large connected-element interferometer.
APPENDIX IIDETAILED BLOCK DIAGRAMS FORTHE GSSR
A more complete explanation of the conceptual diagram
of Fig. 2 is shown in Fig. 15. The ‘‘programmable oscillators’’
or POs of Figure 2 are further explained below. The PO is a
frequency generator whose output signal is controlled by a
Bnumerically controlled oscillator[ (NCO). The NCO is, inturn, driven by an embedded computer, which changes the
output frequency as given by reading an ephemeris file
appropriate to the observation. The next generation PO
(currently being integrated into the system) is a mixed-signal
multilayer printed circuit board (PCB) which uses a field-
programmable gate array (FPGA) in place of the embedded
computer. The ephemeris information is loaded into the
FPGA through a dynamic reconfiguration port. This portgives the ability to reconfigure only that portion of the FPGA
that holds the polynomials for the ephemeris calculation,
while leaving unchanged the portion of the device that
implements the main PO function of calculating the
frequency as a function of time.
Acknowledgment
The authors gratefully acknowledge the improvements
made to this paper based on comments from reviewer
Dr. L. Harcke and several anonymous reviewers.
RE FERENCES
[1] S. J. Ostro, BPlanetary radar,[ in Encyclopediaof the Solar System, L. A. McFadden,P. Weissman, and T. V. Johnson, Eds., 2nd ed.New York: Academic, 2007, pp. 735–764.
[2] R. M. Goldstein, R. R. Green, andH. C. Rumsey, BVenus radar brightnessand altitude images,[ Icarus, vol. 36, no. 3,pp. 334–352, Aug. 1978.
[3] J. K. Harmon, BPlanetary delay-Doppler radarand the long-code method,[ IEEE Trans.Geosci. Remote Sens., vol. 40, no. 9,pp. 1904–1916, Sep. 2002.
[4] L. J. Harcke, BRadar imaging of solar systemices,[ Ph.D. dissertation, Dept. Elect.
Eng., Stanford Univ., Stanford, CA,2005.
[5] N. J. S. Stacy, BHigh-resolution syntheticaperture radar observations of the moon,[Ph.D. dissertation, Dept. Astronomy, CornellUniv., Pittsburgh, PA, 1993.
[6] J. L. H. Webb, D. C. Munson, andN. J. S. Stacy, BHigh-resolution planetaryimaging via spotlight-mode synthetic apertureradar,[ IEEE Trans. Image Process., vol. 7,no. 11, pp. 1571–1582, Nov. 1998.
[7] I. I. Shapiro, S. H. Zisk, A. E. E. Rogers,M. A. Slade, and T. W. Thompson, BLunartopography: Global determination by radar,[Science, vol. 178, no. 4064, pp. 939–948,Dec. 1972.
[8] S. H. Zisk, BA new, earth-based radartechnique for the measurement of lunartopography,[ The Moon, vol. 4, no. 3/4,pp. 296–306, Sep. 1972.
[9] H. A. Zebker and R. M. Goldstein,BTopographic mapping from interferometricSAR observations,[ J. Geophys. Res., vol. 91,no. B5, pp. 4993–4999, Apr. 1986.
[10] R. Gens and J. L. Van Genderen, BReviewarticle: SAR interferometry-issues,techniques, applications,[ Int. J. Remote Sens.,vol. 17, pp. 1803–1835, 1996.
[11] J. L. Margot, D. B. Campbell, R. F. Jurgens,M. A. Slade, and N. J. Stacey, BDigitalelevation models of the moon,[ IEEE Trans.
Slade et al.: Goldstone Solar System Radar Observatory
| Proceedings of the IEEE 11
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
Geosci. Remote Sens., vol. 38, no. 2, pt. 2,pp. 1122–1133, Mar. 2000.
[12] S. Hensley, E. Gurrola, L. Harcke, M. Slade,K. Quirk, M. Srinivasan, C. Lee, S. Yun, J. Jao,B. Wilson, E. De Jong, N. Marechal,L. Weintraub, R. Dickinson, R. Bloom,G. Karamyan, and A. Lilje, BLunartopographic mapping using a new highresolution mode for the GSSR radar,[ in Proc.IEEE Nat. Radar Conf., Washington, DC,May 10–14, 2010, pp. 464–469.
[13] S. Hudson, BThree-dimensionalreconstruction of asteroids from radarobservations,[ Remote Sens. Rev., vol. 8,pp. 195–203, Jul. 1993.
[14] C. Magri, S. J. Ostro, D. J. Scheeres,M. C. Nolan, J. D. Giorgini, L. A. M. Benner,and J.-L. Margot, BRadar observations and aphysical model of asteroid 1580 Betulia,[Icarus, vol. 186, pp. 152–177, Jan. 2007.
[15] C. Magri, SHAPE Software Introduction andDocumentation. [Online]. Available e-mail:[email protected] .
[16] L. A. M. Benner, S. J. Ostro, M. C. Nolan,J.-L. Margot, J. D. Giorgini, R. S. Hudson,R. F. Jurgens, M. A. Slade, E. S. Howell,D. B. Campbell, and D. K. Yeomans, BRadarobservations of asteroid 1999 JM8,[Meteoritics Planet. Sci., vol. 37, pp. 779–792,Jun. 2002.
[17] R. S. Hudson and S. J. Ostro, BShape andnon-principal axis spin state of asteroid 4179Toutatis,[ Science, vol. 270, pp. 84–86,Oct. 1995.
[18] L. A. M. Benner, M. W. Busch, S. J. Ostro,J. D. Giorgini, A. A. Hine, J. K. Harmon,M. C. Nolan, R. Rose, R. F. Jurgens, J. S. Jao,C. Magri, and J.-L. Margot. (2006, Aug.).Radar images of asteroid 100085 (1992 UY4).presented at the IAU Symp. No. 236,Near-Earth Objects, Prague, Czech Republic.[Online]. Available: http://hdl.handle.net/2014/39808.
[19] S. J. Ostro, R. F. Jurgens, K. D. Rosema,R. S. Hudson, J. D. Giorgini, R. Winkler,D. K. Yeomans, D. Choate, R. Rose,M. A. Slade, S. D. Howard, D. J. Scheeres, andD. L. Mitchell, BRadar observations ofasteroid 1620 geographos,[ Icarus, vol. 121,pp. 46–66, May 1996.
[20] S. J. Ostro, R. S. Hudson, R. F. Jurgens,K. D. Rosema, R. Winkler, D. Howard,R. Rose, M. A. Slade, D. K. Yeomans,J. D. Giorgini, D. B. Campbell, P. Perillat,J. F. Chandler, and I. I. Shapiro, BRadarimages of asteroid 4179 Toutatis,[ Science,vol. 270, pp. 80–84, Oct. 1995.
[21] R. S. Hudson, S. J. Ostro, and D. J. Scheeres,BHigh-resolution model of asteroid 4179Toutatis,[ Icarus, vol. 161, pp. 346–355,Feb. 2003.
[22] D. K. Yeomans, S. J. Ostro, and P. W. Chodas,BRadar astrometry of near-Earth asteroids,[Astron. J., vol. 94, pp. 189–200, Jul. 1987.
[23] S. J. Ostro, J. L. Margot, L. A. M. Benner,J. D. Giorgini, D. J. Scheeres,E. G. Fahnestock, S. B. Broschart, J. Bellerose,M. C. Nolan, C. Magri, P. Pravec, P. Scheirich,R. Rose, R. F. Jurgens, E. M. de Jong, andS. Suzuki, BRadar imaging of binarynear-earth asteroid (66391) 1999 KW4,[Science, vol. 314, pp. 1276–1280, Nov. 2006.
[24] C. E. Cook and M. Bernfeld, Radar Signals,1st ed. New York: Academic, 1967.
[25] C. Elachi, Introduction to the Physics andTechniques of Remote Sensing, 1st ed.New York: Wiley, 1987ISBN: 0471848107.
[26] J. L. Margot, Planetary, Radar AstronomyWith Linear FM (Chirp) Waveforms, InternalArecibo Memo, Jun. 2000.
[27] K. Quirk and M. Srinivasan, GSSR Waveformsfor Lunar Observations, Internal JPL Study,Mar. 2008.
[28] NASA. (2010). Near Earth Object Program.[Online]. Available: http://neo.jpl.nasa.gov/news/news167.html.
[29] G. Chin, S. Brylow, M. Foote, J. Garvin,J. Kasper, J. Keller, M. Litvak, I. Mitrofanov,D. Paige, K. Raney, M. Robinson, A. Sanin,D. Smith, H. Spence, P. Spudis, S. A. Stern,and M. Zuber, BLunar reconnaissance orbiteroverview: The instrument suite and mission,[Space Sci. Rev., vol. 129, no. 4, pp. 391–419,Apr. 2007.
[30] I. V. Kholin, BAccuracy ofbody-rotation-parameter measurementwith monochromatic illumination andtwo-element reception,[ Radiophys. QuantumElectron., vol. 35, pp. 284–287, May 1992.
[31] J. L. Margot, S. J. Peale, R. F. Jurgens,M. A. Slade, and I. V. Holin, BLarge longitudelibration of Mercury reveals a molten core,[Science, vol. 316, pp. 710–714, May 2007.
[32] M. A. Slade, S. Zohar, and R. F. Jurgens,BVenus: Improved spin vector fromGoldstone radar observations,[ Astronom. J.,vol. 100, pp. 1369–1374, Oct. 1990,(plates 1396–1401).
[33] I. I. Shapiro, J. F. Chandler, D. B. Campbell,A. A. Hine, and N. J. S. Stacy, BThe spinvector of Venus,[ Astronom. J., vol. 100,pp. 1363–1368, Oct. 1990.
[34] G. S. Downs, P. E. Reichley, and R. R. Green,BRadar measurements of Martian topography
and surface properties: The 1971 and 1973oppositions,[ Icarus, vol. 26, pp. 273–312,Nov. 1975.
[35] G. L. Tyler, D. B. Campbell, G. S. Downs,R. R. Green, and H. J. Moore, BRadarcharacteristics of Viking 1 landing sites,[Science, vol. 193, pp. 812–815, Aug. 1976.
[36] A. F. C. Haldemann, D. L. Mitchell,R. F. Jurgens, M. A. Slade, andD. O. Muhleman, BMars pathfinder landingsite assessment with Goldstone delay-Dopplerand CW-radar experiments,[ J. Geophys. Res.Planets, vol. 102, no. E2, pp. 4097–4106,Feb. 1997.
[37] M. P. Golombek, J. A. Grant, T. J. Parker,D. M. Kass, J. A. Crisp, S. W. Squyres,A. F. C. Haldemann, M. Adler, W. J. Lee,N. T. Bridges, R. E. Arvidson, M. H. Carr,R. L. Kirk, P. C. Knocke, R. B. Roncoli,C. M. Weitz, J. T. Schofield, R. W. Zurek,P. R. Christensen, R. L. Fergason,F. S. Anderson, and J. W. Rice, Jr., BSelectionof the Mars Exploration Rover landing sites,[J. Geophys. Res., vol. 108, pp. 13–33,Dec. 2003.
[38] A. F. C. Haldemann and B. J. Butler,BEvaluation of the Phoenix Region B landingsite rock coverage from available radar data,[in Proc. 4th Int. Conf. Mars Polar Sci. Explorat.,Oct. 2006, p. 8057, LPI ContributionNo. 1323.
[39] D. O. Muhleman, A. W. Grossman,B. J. Butler, and M. A. Slade, BRadarreflectivity of Titan,[ Science, vol. 248,pp. 975–980, May 1990.
[40] D. B. Campbell, G. J. Black, L. M. Carter, andS. J. Ostro, BRadar evidence for liquid surfaceson Titan,[ Science, vol. 302, no. 5644,pp. 431–434, Oct. 2003.
[41] F. C. Vandenbussche. (1999, Mar.). SOHO’sRecoveryVAn Unprecedented Success, ESA,Noordwijk, The Netherlands. [Online].Available: http://www.esa.int/esapub/bulletin/bullet97/vandenbu.pdf.
[42] R. M. Goldstein, S. J. Goldstein, andD. J. Kessler, BRadar observations of spacedebris,[ Planetary Space Sci., vol. 46,pp. 1007–1013, Aug. 1998.
[43] J. Lauf, M. Calhoun, W. Diener, J. Gonzalez,A. Kirk, P. Kuhnle, B. Tucker, C. Kirby, andR. Tjoelker. (2005, Aug.). Clocks andTiming in the NASA Deep Space Network,Jet Propulsion Laboratory, Pasadena, CA.[Online]. Available: http://www.dtic.mil/cgibin/GetTRDoc?Location=U2&doc=GetTRDoc.pdf&AD=ADA483368.
ABOUT THE AUT HORS
Martin A. Slade (Member, IEEE) received the B.S.
and M.S. degrees in physics and the Ph.D. degree
in earth, atmospheric, and planetary science
from the Massachusetts Institute of Technology,
Cambridge in 1964, 1967, and 1971 respectively.
Currently, he is a Group Supervisor of the Solar
System Radar Group at the Jet Propulsion Labo-
ratory (JPL), California Institute of Technology,
Pasadena. His scientific interests have ranged
from testing of gravitational theories that gener-
alize Bmetric[ theories (such as general relativity) to very long baseline
observations of the structure of quasars and active galactic nuclei.
Currently, the determinations of time variations in the spin vector and
rotation period for both Mercury and Venus are ongoing collaborative
research areas in radar astronomy. Other scientific interests are radar
mapping of the topography of the lunar poles and the surface of Mercury.
The determination by radar of the shapes, orbits, and physical
characterization for near-Earth asteroids and comets is an active area
of collaboration with scientists at JPL and elsewhere.
Dr. Slade is a member of the American Geophysical Union, the
American Astronomical Society (Division for Planetary Sciences and
Division on Dynamical Astronomy), and the International Astronomical
Union.
Slade et al. : Goldstone Solar System Radar Observatory
12 Proceedings of the IEEE |
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
Lance A. M. Benner received the A.B. degree in
physics from Cornell University, Ithaca, NY, in 1987
and the Ph.D. degree in earth and planetary
science from Washington University in St. Louis,
St. Louis, MO, in 1994.
Currently, he is a Research Scientist at the Jet
Propulsion Laboratory, California Institute of
Technology, Pasadena. He has served in this
position since 2003. Prior to that, he was a
Scientist from 1998, and a National Research
Council Postdoctoral Fellow from 1995. Both of these positions were at
the JPL. Prior to work at JPL, he was a Postdoctoral Research Associate,
Washington University, Seattle, from 1994. He has authored more than
50 papers on asteroids and comets. He specializes in radar imaging of
near-Earth and main-belt asteroids using the NSF’s Arecibo Observatory
(Puerto Rico) and NASA’s Goldstone Solar System Radar (California), and
he has participated in radar observations of more than 200 near-Earth
asteroids.
Dr. Benner is a member of the American Astronomical Society
(Division for Planetary Sciences), the Meteoritical Society, the American
Geophysical Union, and the International Meteor Organization.
Arnold Silva received the B.S. degree from the
California State University at Los Angeles, Los
Angeles, in 1981 and the M.S. degree from the
California State University at Northridge,
Northridge, in 1987, both in electrical engineering,
with emphasis on power, communications, and
radar.
In 1981, he joined the Transmitter Group of ITT
Gilfillan and was involved in the design of high-
power modulators, power converters, and control
circuits for the company’s major radar programs. From 1989 to 1993, he
was employed as an Engineering Specialist of Whittaker Electronic
Systems engaged in the design, simulation, and analysis of radar
transmitter systems and subsystems, for both classified and unclassified
programs. His designs involved development of very high-power, low
phase noise transmitters (pulsed and CW). In 1993, he joined the Jet
Propulsion Laboratory, California Institute of Technology, Pasadena, as a
Member of the Technical Staff in the Transmitter Engineering Group. In
October 2010, he became a Senior Project Leader at The Aerospace
Corporation, El Segundo, CA. He has designed ultra phase and amplitude
stable high-voltage power subsystems and RF subsystems for the S-band
and X-band Deep Space Network (DSN) klystron-based transmitter
uplinks. These systems ranged in power up to 500 KW CW radiated. In
1998, he assumed the position of Group Supervisor for the Transmitter
Engineering Group. In 2010, he assumed the role of System Engineer for
overall DSN Facilities, Power, and Infrastructure.
Slade et al.: Goldstone Solar System Radar Observatory
| Proceedings of the IEEE 13